Computing and Interpreting Silhouette Scores

A 1-D clustering produced two clusters: C_1 = \{1, 2\} and C_2 = \{8, 9, 10\}.

1. Compute the silhouette value s(i) for the point x = 2 and for the point x = 8. Show a(i) and b(i).
2. A stakeholder says "silhouette is 0.8, so the segmentation is correct." Give two distinct reasons this conclusion is unsafe.
3. Why would silhouette be a poor way to compare a k-means result with a DBSCAN result on data shaped like two concentric rings?